The present invention relates generally to microcellular carbon foam precursors based on foamed gels of polyacrylonitrile (PAN), and more particularly to microcellular carbon foam prepared from such precursors and characterized by a well-interconnected strut morphology, an expanded d(002) X-ray turbostatic spacing, and a uniform distribution of cell sizes with the majority of cells having diameters less than about 10 micrometers (.mu.m).
Microcellular carbon foams having open porosity with cell sizes less than about 100 .mu.m have been successfully used as catalyst supports, absorbents, filters, electrodes and the like. Microcellular carbon foams have been produced from precursors prepared by various manufacturing processes including sol-gel, replication and phase inversion processes. These processes are generally described in the publication "Low-Density Microcellular Materials", R. W. Hopper et al, Report No. UCRL-JC-104935, Lawrence Livermore National Laboratory Preprint, December 1990. The phase inversion process as described in this publication and in U.S. Pat. No. 4,832,881, issued May 23, 1989, involves dissolving polyacrylonitrile and other carbonizable polymers in a suitable solvent such as maleic anhydride, methyl sulfone blends with water or cyclophexanol or norcamphor, dimethyl sulfoxide with ethylene glycol, ethylene carbonate, and dimethylformamide with water, at an elevated temperature and cooling the resulting solutions at a controlled rate to allow the selected polymer to undergo phase inversion and form a porous gel. The solvent is removed from the gel by vacuum sublimation or extraction with supercritical carbon dioxide. The porous gel is then air cured at an elevated temperature and carbonized to form a microcellular carbon foam product.
While previous phase inversion processes for preparing carbon foam precursors such as generally described above have been found to be successful, there are still some shortcomings which detract from the known phase inversion processes. For example, when using solid solvents such as maleic anhydride only relatively small and thin castings (less than about 1 cm in thickness) of the carbon foam precursors can be produced without the carbon foam being highly stressed and subject to cracking due to the brittle nature of maleic anhydride. Also, when a solid solvent such as maleic anhydride freezes, the resulting stress of the gel during the formation thereof imparts a ring defect know as a Liesegang ring structure into the gel. Further, the production of microcellular carbon foam in a relatively low density range of about 30 to 100 mg/cm.sup.3 by the phase inversion of polyacrylonitrile dissolved in solvents such as gamma-butyrolactone, dimethyl sulfoxide, ethylene carbonate and maleic anhydride has not proven to be adequately satisfactory. It was found that this problem in the production of low density carbon foams was due to the fact that while these solvents are capable of dissolving polyacrylonitrile at elevated temperatures they either fail to adequately release the dissolved polymer at lower temperatures or release the polymer as a poorly interconnected ball precipitate which results in a highly friable carbon foam.
It was also found that the preparation of microcellular carbon foams in a density range of about 30 to 1000 mg/cm.sup.3 while providing such carbon foams with well interconnected strut morphology has not been satisfactorily achieved in a reproducible manner by the practice of previously know phase inversion processes, particularly since the carbon foams could be produced in only relatively small sizes or often exhibited poorly interconnected strut morphology which detracted from the use of such carbon foams in many applications such as filters for corrosives liquids. Additionally, it was expected that microcellular carbon foam with a well interconnected strut morphology as previously available would be useful in the fabrication of super capacitors or in lithium battery applications such as providing an intercalation anode structure for secondary lithium batteries or a cathode for primary lithium batteries. However, even the microcellular carbon foams with well-defined strut morphologies as previously provided have not been proven to be adequate for such super capacitor and battery applications since the expanded d(002) X-ray turbostatic spacing of existing microcellular carbon foams often is less than 3.50 angstroms (.ANG.) which is insufficient to provide for repeated intercalation of Li.sup.+ ions without spalling of the carbon gallery. Moreover, any graphitic character of the carbon reduces the d(002) spacing toward that of graphite (3.37 .ANG.) and further aggravates spalling during intercalation.